基于CUDA的RS纠删码性能优化  

CUDA-based Performance Optimization of RS Erasure Coding

作　　者：戴世航 李小勇[1] 

机构地区：[1]上海交通大学信息安全工程学院

出　　处：《微型电脑应用》2016年第1期70-72,共3页Microcomputer Applications

摘　　要：目前分布式存储系统中保证数据可用性的常用方法有多副本技术和纠删码技术。与多副本技术相比,纠删码技术有更高的存储空间利用率,但附加的编码流程不可避免地带来了较高的时间延迟,影响了系统的实时性,限制了纠删码的应用。为了提高纠删码的编码效率,对开源代码库Jerasure提供的RS纠删码进行优化,利用CUDA对其进行加速。实验结果显示,相对于原始算法,该方法将编码速度提高了约20倍,为纠删码技术应用于实时系统提供了可能。At present,the ways used by distributed storage system to ensure the availability of data mainly have multi-replicate and erasure coding technologies.Compared with multi-replicate,erasure coding saves more storage space,however,it takes more time in encoding,which has a bad effect in speed,and limits its application.In order to improve the encoding efficiency of erasure coding,an algorithm provided in the open-source library Jerasure is accelerated by CUDA in this paper.The test result shows that the accelerated algorithm is about 20 times faster than the original one.

关 键 词：RS纠删码 CUDA GPU加速 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]